Catena 207 (2021) 105588

Contents lists available at ScienceDirect

Catena
journal homepage: www.elsevier.com/locate/catena

The interactive effects of elevation, precipitation and lithology on karst

rainfall and runoff erosivity
Yuan Jiang a, b, Jiangbo Gao a, *, Lin Yang c, Shaohong Wu a, b, Erfu Dai d
a
Key Laboratory of Land Surface Pattern and Simulation, Institute of Geographic Sciences and Natural Resources Research, Chinese Academy of Sciences, Beijing, China
b
College of Resources and Environment, University of Chinese Academy of Sciences, Beijing, China
c
School of Geography and Ocean Science, Nanjing University, Nanjing, China
d
Lhasa Plateau Ecosystem Research Station, Key Laboratory of Ecosystem Network Observation and Modeling, Institute of Geographic Sciences and Natural Resources
Research, Chinese Academy of Sciences, Beijing, China

A R T I C L E I N F O A B S T R A C T

Keywords: Soil erosion, which is the prominent ecological and environmental problem in the karst regions of southwestern
Optimized erosivity model China, has seriously affected the regionecological conservation and economic development. Quantifying rainfall
Spatial differentiation erosivity and runoff erosivity, can be a means to identify the dominant environmental factors involved in these
Dynamic changes
mechanisms and provide a theoretical basis for mitigating soil and water loss. This aids determination of the
Driving factors
measures required to control rocky desertification and promote the ecological restoration of karst areas.
Geographical detector
Accordingly, this study simulated the rainfall erosivity in the Sancha River Basin (SRB) between 1985 and 2014
using an adopted daily rainfall model according to the karst erosive rainfall standard. Based on the degree of
rocky desertification in the karst areas, correction coefficients were applied to improve runoff erosivity calcu­
lation and estimate the runoff erosivity of the basin between 1993 and 2014. In addition, the dominant and
interactive factors affecting rainfall/runoff erosivity of the diverse geomorphological types in the SRB were
quantitatively identified using the geographical detector. Results showed that the rainfall erosivity between 1985
and 2014 averaged 6913.73 MJ⋅mm⋅ha− 1⋅h− 1, with an average runoff erosivity of 1121.37 m4⋅km− 2⋅s− 1 between
1993 and 2014. Precipitation and elevation were found to be the dominant factors shaping the spatial distri­
butions of rainfall/runoff erosivity in the SRB. Precipitation explained over 90% of the spatial distribution in
rainfall erosivity, and the q value of precipitation for runoff erosivity had no direct relationship with time.
Furthermore, the interactions of elevation and precipitation, elevation and lithology type had prominent effects
on rainfall erosivity and runoff erosivity in the SRB, respectively. Among them, the compound effects of elevation
and precipitation could explain more than 80% of rainfall erosivity. These findings should be essential for
managing soil and water loss in the karst areas.

1. Introduction erosion. Karst rainfall erosivity characterizes the impact of precipitation

The karst ecosystem develops through atmosphere–water–rock–bi­
ology interactions based on carbonate rock, and has a unique compo­
sition, structure, and function. Due to the distinct geochemical processes
in the karst areas of the Guizhou Province, the thickness of soil layer in
this region is generally thin (Wang et al., 2019). Under equal erosion
intensity, the damage caused by soil erosion in the region is far greater
than in non-karst areas with thicker soil layers (Zeng et al., 2017).
Considering the extensive potential damage caused by erosion, rainfall
and runoff are determined as the main driving factors leading to soil
on soil erosion in karst areas, and is mainly affected by rainfall volume,
intensity, pattern, and frequency. Karst runoff erosivity is one of the
major driving forces of soil erosion processes on slopes or in watersheds
and can be responsible for the erosion of slopes, gullies and valleys in
karst areas by dislodging soil and transporting sediment. One approach
for quantifying such damage is the universal soil loss equation (USLE)
model, in which the rainfall factor of soil erosion is expressed as rainfall
erosivity (Jelinski and Yoo, 2016), and estimation of sediment loss on
the slope scale is characterized by runoff erosivity. Thus, accurate as­
sessments of rainfall erosivity, runoff erosivity, and identification of the

* Corresponding author.
E-mail addresses: jiangy.18s@igsnrr.ac.cn (Y. Jiang), gaojiangbo@igsnrr.ac.cn (J. Gao), yanglin@nju.edu.cn (L. Yang), wush@igsnrr.ac.cn (S. Wu), daief@igsnrr.
ac.cn (E. Dai).

https://doi.org/10.1016/j.catena.2021.105588
Received 11 March 2021; Received in revised form 6 July 2021; Accepted 7 July 2021
Available online 17 July 2021
0341-8162/© 2021 Elsevier B.V. All rights reserved.
Y. Jiang et al.
precipitation on the middle elevation terrace exceeding 0.55 (Fig. S5b).
For the middle elevation hill, the change in the explanatory power of the
various environmental factors for runoff erosivity variability had no
direct relationships with time. In relatively flat areas, the combination of
precipitation and lithology type decreased with increasing in terrain
relief between 1993 and 2014, in the order of middle elevation plain >
middle elevation terrace (Tables S17–S21).
4. Discussion
For this study, we optimized the model of rainfall erosivity based on
the karst erosive rainfall standard and the model of runoff erosivity
according to bedrock bareness rate. Moreover, the dominant and
interactive factors of rainfall/runoff erosivity and variability in the SRB
and diverse geomorphological types were quantitatively identified, to
provide relevant policy recommendations for controlling the soil and
water loss in karst areas.
4.1. Dominant factors of karst rainfall/runoff erosivity
This study confirmed that precipitation, elevation, vegetation
coverage, lithology type, land use type and slope all have important
effects on rainfall/runoff erosivity and variability in the SRB and
different geomorphological types. In particular, precipitation and
elevation were clearly significant for determining the spatial distribu­
tion of rainfall/runoff erosivity, which precipitation is a prerequisite for
rainfall erosivity and an indispensable source of water for surface runoff
mechanisms (Xin et al., 2011). Meanwhile, the high relief terrain causes
the adiabatic expansion and cooling of rising air masses, which in turn
increases relative humidity, leading to cloud formation and precipita­
tion that ultimately affect rainfall/runoff erosivity (Meshesha et al.,
2015).
The interaction of environmental factors was found to enhance the
explanatory power of rainfall/runoff erosivity. In particular, the com­
binations of precipitation with various other environmental factors had
the strongest interactive impact on the rainfall erosivity. For the middle
elevation plain, the interaction between vegetation coverage and pre­
cipitation had greater explanatory power for the spatial distribution of
rainfall erosivity than any single factor. This is because vegetation can
effectively reduce the kinetic energy of rainfall and prevent the soil
surface from being directly impacted by raindrops, thus leading to
promotion of rain infiltration, and reduction runoff energy (Xu et al.,
2018). The pairwise combinations of precipitation and lithology type
(1993–1999), precipitation and elevation (2000–2014) were domi­
nantly influenced the distribution of runoff erosivity in the middle relief
mountain. This is because lithological conditions in watersheds often
determine the infiltration and evaporation of water and the maximum
underground water storage capacity. Moreover, these factors also affect
the spatial distribution of evaporation, infiltration, and groundwater (Li
et al., 2019).
4.2. Characteristics of karst rainfall/runoff erosivity variability
In the SRB, climate change and geological conditions have resulted in
pronounced changes in rainfall/runoff erosivity (Amanambu et al.,
2019). The results of this study revealed that precipitation, elevation,
and lithology type were the main driving factors of rainfall/runoff
erosivity variability in the karst areas. Although the combined effect of
precipitation and elevation had the greatest initial impact on rainfall
erosivity (1995–2009), the interaction between precipitation and
geomorphological types showed dominance in 2010–2014, demon­
strating the growing influence of terrain relief. The explanatory power of
vegetation coverage for spatial distribution on runoff erosivity vari­
ability declined from 1993 to 2014. In recent years, the NDVI value of
SRB has been steadily increasing, with the value of 0.685 and 0.740 in
2000 and 2005, respectively. In 2010, the NDVI of SRB was 0.762 and
Catena 207 (2021) 105588
value in 2014 was 0.824. With the advancement of ecological man­
agement projects in China’s karst areas including the ChangZhi Project
and the Natural Forest Protection Project, the overall ecological and
vegetation conditions in the karst areas are improving significantly,
which may explain the decrease in q values for vegetation coverage in
runoff erosivity variability over time. The interaction between precipi­
tation and elevation dominated runoff erosivity variability, with the q
values exceeding 0.15. The middle elevation hill was more sensitive to
changes in elevation, which agrees with previous research findings
wherein the stratification of elevation was found to reflect comprehen­
sive differences in climate, vegetation, and topography related to the
geomorphological type (Kotlarski et al., 2015).
4.3. Uncertainty analysis and future perspectives
The average annual rainfall erosivity in the SRB for 1985–2014
ranged from 4351.45 to 8632.20 MJ⋅mm⋅ha− 1⋅h− 1, with the average
value of 6913.73 MJ⋅mm⋅ha− 1⋅h− 1. In addition, Xu et al. (2005) used a
monthly rainfall erosivity model to calculate the rainfall erosivity in the
Guizhou Province, where the average rainfall erosivity for 1951–2001
was 4383.34 MJ⋅mm⋅ha− 1⋅h− 1. Zhu et al. (2019) also simulated rainfall
erosivity in the Guizhou Province for the period between 1960 and
2017, and obtained average values of 5825.60 MJ⋅mm⋅ha− 1⋅h− 1. The
average value of rainfall erosivity found in this study was higher than
the previous studies because the Guizhou Province includes karst and
non-karst areas. The rainfall erosivity of non-karst areas is lower than
that of karst areas, and the calculation methods of previous studies for
rainfall erosivity did not distinguish in karst and non-karst areas. This
study promoted the rainfall erosivity model, to calculate rainfall
erosivity in karst areas more accurately.
Runoff is one of the main driving forces of soil erosion, which erodes
surfaces, slopes and valleys by dislodging and separating soil and initi­
ating sediment transport (Dai et al., 2018). Currently, related research is
mostly limited to empirical inferential statistics, which generates un­
certainty in the estimation of runoff erosivity. In this study, the SWAT
model was used to simulate changes to runoff erosivity in the SRB. The
model had good accuracy and applicability, making it a strong foun­
dation for future studies on runoff erosivity in the SRB and similar karst
areas. However, the SWAT method has some limitations that merit
further investigation. For a hydrological model, minimizing uncertainty
is critical for the verification of model accuracy. Therefore, future work
should focus on expanding the number of iterations of the SWAT model
to reduce uncertainty and improve the accuracy of model calibration
and validation.
Rainfall/runoff erosivity and variability were affected by various
environmental factors that may affect the spatial distribution of
erosivity. In this study, natural environmental factors were analyzed to
identify the dominant and interactive factors affecting rainfall/runoff
erosivity and variability. As the geographical detector is based on
stratified variance, it can eliminate collinearity between independent
variables and effectively overcome the limitations of traditional statis­
tical analysis in processing categorical variables.
5. Conclusion
In this study, long-term rainfall erosivity (1985–2014) and runoff
erosivity (1993–2014) in the karst basin of the SRB in the Guizhou
Province, southwestern China were evaluated. Modified rainfall
erosivity and runoff erosivity calculation were applied in this study,
along with the geographical detector, which was employed to quanti­
tatively identify the dominant and interactive factors influencing rain­
fall/runoff erosivity in the SRB according to diverse geomorphological
types. The main findings are as follows.
The average rainfall erosivity in the SRB from 1985 to 2014 was
6913.73 MJ⋅mm⋅ha− 1⋅h− 1, while the average runoff erosivity from 1993
to 2014 was 1121.37 m4⋅km2⋅s− 1. Both rainfall erosivity and runoff